Finding absolute minimum and maximum using lagrange mutlipliers?

58 Views Asked by At

I have this function

$f(x,y)=3y^2-12x^2+1$ with the constraint $h(x,y)=x^2+y^2-x-2y+\frac54=0$.

First thing I do is write

$$3y^2-12x^2+1-λ(x^2+y^2-x-2y+\frac54)$$ which expands to $$3y^2-12x^2+1-λx^2-λy^2+λx+2yλ-\frac54.$$

The partial derivative with respect to $x$:

$-24x-2λx+λ=0$ where I get that $x= \dfrac λ{2(12+λ)}$.

The partial derivative with respect to $y$:

$6y-2yλ+2λ=0$ where I get that $y= -\dfrac λ{3-λ}$.

The partial derivative with respect to $λ$:

$$-x^2-y^2+x+2y$$

Now I know that I need to plug in the values of $x$ and $y$ in the $λ$ partial derivative which is $-x^2-y^2+x+2y$.

And here is where my problems arise.

On the answer sheet it states that $y= \dfrac λ{3-λ}$ (different from my answer since I get a negative $λ$ on the numerator).

Consequently, it states that I need to plug in the values of $x$ and $y$ NOT in the partial derivative of $λ$ (which I thought was always the case) but it states that I need to plug the values of $x$ and $y$ in the CONSTRAINT, hence $x^2+y^2-x-2y+\frac54$.

So my two questions are :

1) Why am I getting a different $y$ value?

2) Who is in the wrong here? Do I need to plug the $x$ and $y$ values in the partial derivative of $λ$ or in the constraint formula? Why?

Thanks in advance for the help. I have some learning disabilities involving maths so I would really appreciate your answers being as simple and clear as possible. Thanks!!

2

There are 2 best solutions below

6
On BEST ANSWER

The $\lambda$ partial derivative should have the $5/4$ still in it. The constraint and $dL/d\lambda$ should be the same.
I usually add $\lambda$ times the constraint instead of subtracting it so my $\lambda$ would be the negative of yours.

4
On

Hint.

Your constraint

$$ x^2 + y^2 - x - 2 y + 5/4 = 0 $$

defines a feasible point region at $x = \frac 12, y = 1$

so the minimum is $f(\frac 12,1) = 1$

NOTE

In this case you cannot apply the Lagrange Multipliers technique because it needs that the restriction should define a derivable real variety, which is not the case.